An approach to automatic model generation for stochastic qualitative simulation of building air conditioning systems

The authors propose a stochastic qualitative simulation which can derive behavior roughly from a simple model of a target system. In this method, the model must be constructed with a number of stochastic parameters. The parameter tuning process is the most difficult work in the model construction. This paper proposes a method of generating qualitative models based on automatic parameter turning. In this method, a tentative model is selected from existing models and is modified in advance. Parameters of functions or propagation rules are then optimized.